Isolation and characterization of new secondary metabolites from Asphodelus microcarpus

Abstract

Phytochemical study of the ethanolic extract of Asphodelus microcarpus Salzm. et Viv. (Asphodelaceae) resulted in the isolation of two new compounds, methyl-1,4,5-trihydroxy-7-methyl-9,10-dioxo-9,10-dihydroanthracene-2-carboxylate (1), and (1R) 3,10-dimethoxy-5-methyl-1H-1,4-epoxybenzo[h]isochromene (2) as well as three known compounds; 3,4-dihydroxy-methyl benzoate (3), 3,4-dihydroxybenzoic acid (4), and 6-methoxychrysophanol (5). Compound 1 showed a potent activity against methicillin-resistant Staphylococcus aureus (MRSA) and Staphylococcus aureus with IC₅₀ values of 1.5 and 1.2 µg/mL, respectively. Compound 3 showed antileishmanial activity with an IC₅₀ value of 33.2 µg/mL. Compound 2 is the first isochromene possessing a highly strained 1,4-epoxy moiety. The structure elucidation of isolated metabolites was carried out using spectroscopic data, the absolute configuration of 2 based on optical rotation and electronic circular dichroism experiments and calculations.

Introduction

The genus Asphodelus Reichb. (Asphodelaceae) is a circum-Mediterranean genus, which includes five sections and is represented by 16 species (Lifante and Aguinagalde, 1996). Asphodelus microcarpus Salzm. et Viv. (Asphodelaceae) is a stout robust herb with roots of several spindle-shaped tubers, widely distributed over the coastal Mediterranean region (Tackholm, 1974). A. microcarpus is used to treat ectodermal parasites, jaundice, and psoriasis, and is also used by Bedouins as an antimicrobial agent (Tackholm, 1974). A literature survey revealed that lipids, sterols, triterpenes, anthraquinones, and arylcoumarins have been isolated from A. microcarpus (El-Seedi, 2007; Ghoneim et al., 2013). Anthraquinones and pre-anthraquinones are considered to be important chemotaxonomic markers for plants in the family Asphodelaceae (Van et al., 1995). In continuation of our effort for searching new antimicrobial metabolites from natural sources (Ghoneim et al., 2013). Five compounds (1–5) were isolated from the tubers of A. microcarpus, and their antimicrobial activities were evaluated.

Materials and methods

General experimental procedures

The optical rotation of 2 was determined with an Autopol IV instrument at room temperature. IR spectra were obtained using a Bruker Tensor 27 instrument. CD spectrum was measured on a JASCO J-715 spectrometer. NMR spectra were recorded on a Bruker Avance DRX-500 instrument at 500 (¹H) and 125 MHz (¹³C), and a Varian Mercury 400 MHz spectrometer at 400 (¹H) and 100 MHz (¹³C). The HRESIMS spectra were measured using a Bruker Bioapex-FTMS with electrospray ionization (ESI). Column chromatographic separation was performed on silica gel 60 (0.04–0.063 mm) and Sephadex LH-20 (0.25–0.1 mm, Merck). TLC was performed on precoated TLC plates with silica gel 60 F254 (0.2 mm, Merck). Semipreparative HPLC (Waters Delta Prep 4000) was performed using Luna^® RP-18 (250 mm, 10 mm, 5 µm). The solvent systems used for TLC analyses were: EtOAc:n-hexane (7:3), CHCl₃:MeOH (9.5:0.5), and CHCl₃:MeOH (8:2).

Plant material

The tubers of A. microcarpus were collected from an area 70 km West of Marsa Matrouh, Egypt, during March 2011. The plant was authenticated by Dr. Ibrahim El-Garf, Professor of Plant Taxonomy, Cairo University, Egypt. A voucher specimen (AM 21) has been deposited in the Pharmacognosy Department, Faculty of Pharmacy, Al-Azhar University, Cairo, Egypt.

Antileishmanial assay

The antileishmanial activity of the isolated metabolites was tested in vitro against a culture of L. donovani promastigotes (S1 strain) grown in RPMI 1640 medium supplemented with 10 % GIBCO fetal calf serum at 26 °C. A 3-day old culture was diluted to 5 × 10⁵ promastigotes/mL, drug dilutions (50–3.1 µg/mL) were prepared directly in cell suspension in a 96-well plate, followed by incubation (26 °C, 48 h). Growth of leishmanial promastigotes was determined by the Alamar Blue assay (BioSource International, Camarillo, CA). Standard fluorescence was measured by a Fluostar Galaxy plate reader (excitation wavelength, 544 nm; emission wavelength, 590 nm). Pentamidine (IC₅₀ 1.01 and IC₉₀ 2.03 µg/mL) and amphotericin B (IC₅₀ 0.47 and IC₉₀ 0.65) were used as the drug controls. Percent growth was calculated and plotted against the tested concentrations in order to determine the IC₅₀ and IC₉₀ values (Ma et al., 2004).

Antimicrobial assay

Compounds 1–5 were tested for antimicrobial activity against Staphylococcus aureus ATCC 29,213, methicillin-resistant S. aureus ATCC 33591 (MRSA), Escherichia coli ATCC 35218, Pseudomonas aeruginosa ATCC 27853, Mycobacterium intracellulare ATCC 23068, Candida albicans ATCC 90028, Candida glabrata ATCC 90030, Candida krusei ATCC 6258, Cryptococcus neoformans ATCC 90113, and Aspergillus fumigatus ATCC 204305 (Bharate et al., 2007; Ma et al., 2004). Ciprofloxacin and amphotericin B were used as positive controls for bacteria and fungi, respectively.

Chemistry

Extraction and isolation

The air-dried powdered plant material (2 kg) was exhaustively extracted by maceration with 70 % EtOH (10 L × 3) at room temperature. The combined ethanolic extracts were concentrated under reduced pressure to afford 400 g residue. The residue was successively partitioned with petroleum ether, EtOAc, and MeOH. Each extract was evaporated to yield 20 g (Pet. ether), 210 g (EtOAc), and 15 g (MeOH) of residue. The MeOH fraction (15 g) was subjected to vacuum liquid chromatography (VLC) on silica gel (600 g) using 1.0 L each of n-hexane, n-hexane/EtOAc (50:50) and (25:75), EtOAc, EtOAc/MeOH (1:1), and MeOH to give six fractions (F1–F6). Fraction F1 (2.1 g) was subjected to VLC on Silica gel (100 g) to deliver eight subfractions (1–8). Subfraction 7 (100 mg) was subjected to subsequent purification on semi-preparative RP-C18 HPLC eluting with MeOH/H₂O (75:25) using Luna^® RP-18 (250 mm, 10 mm, 5 µ) at a flow rate of 5.0 mL/min and detection at λ_max = 254 nm to give compound 2 (2 mg). Fraction F2 (6 g) was subjected to VLC on silica gel (200 g) eluting with CHCl₃/MeOH, isocratic system (9:1) to afford 20 subfractions. Subfraction 2 (120 mg) was chromatographed (silica gel, 7 g, n-hexane/EtOAc, 1:1) to give compound 5 (10 mg). Subfraction 18 (50 mg) was purified on Sephadex LH-20, (MeOH/H₂O, 9:1) to yield compound 1 (15 mg). Subfraction 20 (100 mg) was subjected to subsequent purification on semipreparative RP-C₁₈ HPLC eluting with MeOH/H₂O (60:40 v/v) isocratic, using Luna^® RP-18 (250 mm, 10 mm, 5 µ) at a flow rate of 5.0 mL/min and detection at 254 nm to give compounds 3 (10 mg) and 4 (7 mg), respectively.

Methyl-1,4,5-trihydroxy-7-methyl-9,10-dioxo-9,10-dihydroanthracene-2-carboxylate (1)

Reddish amorphous powder; IR (KBr) υ_max 2,954, 1,720, 1,628, 1,446, 1,384, 1,251, 778 cm⁻¹; UV (MeOH) λ_max (log ε): 209 (4.51), 230 (4.10), 289 (4.03), 439 (3.82) nm. HRESIMS [M + H] ⁺ at m/z 329.0656 (calcd. for C₁₇H₁₃O₇, 329.0661). ¹H NMR and ¹³C NMR data are shown in Table 1.

Table 1.

¹H and ¹³C-NMR and HMBC spectral data (400 MHz) for compound 1 in DMSO-d₆ and ¹H and ¹³C-NMR (500 MHz) for compound 2 in CDCl₃

Position	1			Position	2
	δC	δH	HMBC		δC	δH
1	162.3	–		1	89.3	5.81(1H, s)
2	117.6	–		3	169.9	–
3	108.0	6.70 (1H, s)	C-1, C-11, C-4a	4	162.7	–
4	164.1	–		5	132.5	–
5	161.2	–		6	126.2	7.39 (1H, s)
6	124.3	7.12 (1H, d, J = 1.6)	C-8, C-5a,	7	119.2	7.32 (1H, d, J = 8.5)
7	148.5	–		8	129.4	7.52 (1H, t, J = 8)
8	120.7	7.42 (1H, d, J = 1.6)	C-6, C-5a, C-9	9	106.8	6.95 (1H, d, J = 8)
9	181.1	–		10	157.7	–
10	189.7	–		1a	132.5	–
11	166.7	–		4a	111.2	–
1a	132.4	–		10a	113.7	–
4a	108.4	–		6a	136.9	–
5a	113.1	–		Ar-CH3	24.1	2.76 (3H, s)
8a	131.4	–		OCH3	55.9	4.00 (3H, s)
Ar-CH3	21.5	2.35 (3H, s)	C-6, C-7, C-8	OCH3	56.3	4.09 (3H, s)
OCH3	52.3	3.82 (3H, s)	C-11
OH C-4	12.33		C-3, C-4
OH C-5	11.84		C-5

^aAssignment was confirmed by DEPT, HMQC, ¹H NMR and ¹³C NMR (δ in ppm, J values in Hz) experiments

(1R) 3,10-dimethoxy-5-methyl-1H-1,4-epoxybenzo[h]isochromene (2)

Dark brownish amorphous solid; ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}+290$ (c = 0.02, MeOH; IR (KBr) υ_max 2,349, 1,715, 1,261, 1,020, 802 cm⁻¹; UV (MeOH) λ_max (log ε): 195 (3.93), 270 (3.01), 360 (2.40) nm); CD (c = 0.05, MeOH) λ_max (Δε) 363 (−18), 273 (−22), 212 (−33) HRESIMS [M+Na]⁺ at m/z 293.0749 (calcd. for C₁₆H₁₄O₄Na, 293.0790) and [M+H]⁺ ion at m/z 271.0930 (calcd. for C₁₆H₁₅O₄, 271.0970). ¹H NMR and ¹³C NMR data are shown in Table 1.

Results and discussion

The ethanolic extract of A. microcarpus yielded 5 compounds (Fig. 1). Methyl-1,4,5-trihydroxy-7-methyl-9,10-dioxo-9,10-dihydroanthracene-2-carboxylate (1) was obtained as a reddish amorphous powder. HRESIMS gave an [M + H]⁺ ion at m/z 329.0656 (calcd. for C₁₇H₁₃O₇, 329.0661), consistent with the molecular formula of C₁₇H₁₂O₇. The ¹³C NMR, DEPT, and HMQC of 1 (Table 1) displayed 17 carbon signals, including one methyl at δ 21.5 (on C-7), one methoxyl group at δ 52.3, three methines at δ 108.0 (C-3), 124.3 (C-6), and 120.7 (C-8) and nine quaternary carbons at δ 162.3 (C-1), 117.6 (C-2), 164.1 (C-4), 161.2 (C-5), 148.5 (C-7), 132.4 (C-1a), 108.4 (C-4a), 113.1 (C-5a) and 131.4 (C-8a) as well as three carbonyl carbons at δ 181.1 (C-9), 189.7 (C-10), and 166.7 (C-11). The ¹H NMR spectrum showed the presence of two highly deshielded singlets resonating at δ 12.33 and 11.84 due to the presence of two chelated hydroxyl groups on C-4 and C-5, respectively, (Hernandez-Medel et al., 1999). The ¹H NMR spectrum displayed a methyl singlet at δ 2.35 (3H, s), one methoxy singlet at δ 3.82 (3H, s), three aromatic protons at δ 6.70 (1H, s, H-3), 7.12 (1H, d, J = 1.6 Hz, H-6), and 7.42 (1H, d, J = 1.6 Hz, H-8). The HMBC spectrum exhibited cross-peaks from H-3 to C-1, C-4, C-4a, and C-11, from the C-7 methyl protons to C-6, C-7, and C-8, from H-6 to C-5a, C-8, from H-8 to C-5a, C-6, and C-9. Accordingly compound 1 was methyl-1,4,5-trihydroxy-7-methyl-9,10-dioxo-9,10-dihydroanthracene-2-carboxylate.

Fig. 1.

Compounds 1–5

(1R) 3,10-Dimethoxy-5-methyl-1H-1,4-epoxybenzo[h]isochromene (2) was obtained as a dark brownish amorphous solid. HRESIMS gave an [M+Na]⁺ ion at m/z 293.0749 (calcd. for C₁₆H₁₄O₄Na, 293.0790) and [M + H]⁺ ion at m/z 271.0930 (calcd. for C₁₆H₁₅O₄, 271.0970). The ¹³C NMR, DEPT, and HMQC analyses of 2 (Table 1) displayed the presence of 16 carbon signals, including one methyl at δ 24.1 (Me on C-5); two methoxyl groups at δ 55.9 and 56.3; five methines at δ 89.3 (C-1), 126.2 (C-6), 119.2 (C-7), 129.4 (C-8), and 106.8 (C-9); and eight quaternary carbons at δ 169.9 (C-3), 162.7 (C-4), 132.5 (C-5), 157.7 (C-10), 132.5 (C-10b), 111.2 (C-4a), 113.7 (C-10a) and 136.9 (C-6a). The ¹H NMR spectrum displayed one methyl singlet at δ 2.76 (3H, s, Me on C-5); two methoxyl singlets at δ 4.00 (3H, s) and 4.09 (3H, s); four aromatic protons at δ 7.39 (1H, s, H-6), 7.32 (1H, d, J = 8.5 Hz, H-7), 7.52 (1H, t, J = 8 Hz, H-8), and 6.95 (1H, d, J = 8 Hz, H-9); as well as an oxymethine singlet at δ 5.81 (1H, s, H-1).

The HMBC spectrum (Fig. 2) exhibited cross-peaks from H-6 to C-7, C-4a, C-10a, and methyl carbons on C-5; and correlations from methyl protons on C-5 to C-6 and C-4a, indicating that the quaternary C-5 was substituted with one methyl group. Also there were correlations from H-1 to C-3 and C-4a; from H-7 to C-6, C-9, C-10a, and C-6a; from H-8 to C-6a and C-10; from H-9 to C-10a and C-7; the methoxy protons at δ 4.00 showed correlation to C-3 and at δ 4.09 to C-10 indicating the position of the methoxyl groups. Furthermore, the downfield shifts of C-1 and C-3 were caused by the oxygen substitutions as both of them attached to two oxygen atoms. NOESY experiment revealed a correlation between H-1 and the methoxy group on C-3. To determine the absolute configuration of 2, the CD spectrum and optical rotation were measured and compared to calculated results for the R enantiomer. Experimental optical rotation (OR) data showed that 2 had a specific rotation of ${[\mathrm{\alpha }]}_{\mathrm{D}}^{25}+290$ (c = 0.02, MeOH). A conformational search using the OPLS2005 force field (Banks et al., 2005) with Macromodel software resulted in 2 conformers, whereas the MMFF94s force field (Halgren, 1999a, b) gave 36 conformers. Meta-sampling with a stochastic molecular dynamics and simulated annealing cycle (Kirkpatrick et al., 1983; York et al., 1993; Zheng et al., 2009) of the conformers generated by OPLS2005 and MMFF94s, again using Macromodel, resulted in only 2 distinct conformers. We optimized the two conformers using the PM3 (Stewart, 1989) and PM6 (Stewart, 2007) quantum mechanical semi-empirical methods in Gaussian 09 (Frisch et al., 2009). The conformers maintained all the bonds in the meta-sampling and semi-empirical optimizations, during which we did not use any kind of geometrical constraints. However, when we tried to optimize the 38 force-field generated conformers with ab initio methods, as described below, the C1–O2 bond broke in each case. In order to match the calculations to the experimental findings, we added a dynamical constraint to the C1–O2 bond to maintain it to be close to 1.43Å. The ab initio optimizations were carried out with Gaussian 09 (Frisch et al., 2009) using the hybrid density functional theory (DFT) (Ghosh et al., 1984) functionals B3LYP (Vosko et al., 1980) and B3PW91 (Karamanis et al., 2006) and the Coulomb-attenuating hybrid functional CAM-B3LYP (Stewart, 2007) starting from the 38 conformers described above, using the 6-31G** basis set for B3LYP and B3PW91 and the aug-cc-pvtz basis set (Yanai et al., 2004) for all three functionals, and solvent effects were included using the Polarizable Continuum Model (PCM) (Tomasi et al., 2005) CHCl₃ model. The CAM-B3LYP optimizations yielded only 2 distinct conformers with Boltzmann weights of 72.5 and 27.5 %. We obtained 12 conformers with each of the other DFT methods. The CAM-B3LYP/aug-cc-pvtz optimized geometry including the single constraint resulted in a final C1–O2 bond length of ~1.49 Å for each of the two obtained conformers. The total OR was calculated based on the energy-weighted Boltzmann average of each of the low-energy conformers, using CAM-B3LYP/aug-cc-pvtz (with Gaussian 09) (Table 2). The OR of the R enantiomer was calculated to be +262, which is in excellent agreement with the experimental data (Table 2) (considering that the OR of the S enantiomer would hence be calculated as −262, which obviously is far from the experimental value). The experimental ECD spectrum showed a negative Cotton effect with three peaks at 212, 273, and 363 nm (Fig. 3). The calculated ECD spectrum of the R enantiomer, obtained using time-dependent DFT (Sang et al., 2012) (TDDFT) with CAM-B3LYP, matched well with the experimental spectrum (Fig. 3), whereas that of the S enantiomer demonstrated the opposite behavior (data not shown). The CAM-B3LYP simulated ECD spectrum, generated from 160 excited states using Gaussian band shapes for the peaks, had peaks at 229, 290, and 374 nm, similar to the experimental peaks. A more detailed analysis of the stability of the novel structure has been carried out and will be published in due course. The calculated OR and ECD of the R enantiomer both showed excellent agreement with the experimental data. Therefore, we conclude that the structure of 2 is (1R) 3,10-dimethoxy-5-methyl-1H-1,4-epoxybenzo[h]isochromene.

Fig. 2.

Key HMBC and COSY correlations of 2

Table 2.

Optical rotation contributions of the two lowest-energy conformers of the R enantiomer

Conf.	Boltzmann weight (%)	OR	Cumulative Boltzmann- weighted average OR	Experimental OR
1	72.5	325	325
2	27.5	96	262	290

OR is optical rotation

Fig. 3.

Experimental (black) spectrum compared to the calculated (blue) ECD spectrum of the R enantiomer (2). The CAM-B3LYP calculated results were generated from 160 excited states. For the calculated ECD spectrum, the energy-weighted Boltzmann average for the two conformers is shown. The units of molar ellipticity are deg cm² dmol (Color figure online)

Known compounds 3–5 were identified by ¹H NMR, ¹³C NMR, DEPT, HMQC, HMBC, and HRESIMS to be 3, 4-dihydroxy-methyl benzoate 3 (Si et al., 2006), 3, 4-dihydroxybenzoic acid 4 (Núñez-Sellés et al., 2002), and 6-methoxy chrysophanol 5 (Jain et al., 2013). Compounds 1–5 were evaluated for their antimicrobial and antileishmanial activities. Compound 1 showed potent activity against MRSA and S. aureus with IC₅₀ values of 1.5 and 1.2 µg/mL, respectively. Compound 3 showed antileishmanial activity with an IC₅₀ value of 33.2 µg/mL.